This repository implements the first version of the SignON ELAN Processing Framework, which objective is to extract data of interest from ELAN files to be used for Neural Machine Translation and Sign Language Recognition. This framework takes corpora compiled under the ELAN annnotation standard as input and process them to extract the different annotations. Also it allows processing videos to generate clips with the frames in which annotations are generated.
In the file "test" of this repository you can find an example of usage.
This research was developed using Python 3.8.0. Below, the library requirements are listed to assure the experiments reproducibility.
| Resource | Version/URL |
|---|---|
| Numpy | 1.19.5 |
| OpenCV | 4.5.5 |